NetherlandsLogistics and transport2021 – ongoingProRail
AI detection of faulty railway switch heaters
ProRail cut deployment time for switch-heater detection models from months to hours, according to UbiOps
Problem
Railway switches depended on functioning heaters during freezing weather. Deploying and maintaining detection models also required substantial engineering work.
Approach
ProRail's asset specialists and data scientists worked with UbiOps on an AI algorithm that combined sensor measurements with weather data. UbiOps hosted the detection pipeline and connected its data sources. The platform also supported model updates and operational monitoring.
Results
Takeaway
The reported improvement concerned model deployment effort. ProRail did not disclose a measured reduction in switch failures.
Categories
- Industry
- Logistics and transport
- Business function
- Production and qualityIT and R&D
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
UbiOps described an operational AI pipeline for detecting faulty switch heaters. It reported that model deployment took hours instead of months and that the overall system was running within days.
Full details are in the original sources.
